An intake port for guiding air to a cylinder chamber is formed in a cylinder head of a direct injection type diesel engine, and an intake valve arranged in the intake port is opened or closed in response to each stroke of the engine.
Air guided from the intake port to the cylinder chamber is compressed, is mixed with fuel injected from an injection nozzle, and is then combusted. As is well known in this case, as air and fuel are mixed better, a combustion efficiency is improved.
Various means for improving the mixing state of air and fuel have been conventionally used. For example, a technique using a high swirl port (forcible swirl intake hole), a so-called HSP structure, is known.
The HSP structure is as shown in FIGS. 1, 2A, and 2B. Reference numeral 1 denotes a cylinder liner; 2, a cylinder chamber; 3, a cylinder head; and 4, an intake apparatus. Intake apparatus 4 is constituted by intake port 5 and intake valve 6. Reference numeral 7 denotes an exhaust port. A fuel injection nozzle (not shown) is arranged on cylinder head 3 to face cylinder chamber 2.
Intake port 5 is slightly eccentric to the center of intake valve 6. Upon intake stroke wherein intake valve 6 is moved downward and intake port 5 is opened, intake air which receives "eccentricity" due to intake port 5 is guided to cylinder chamber 2, and a swirl is focibly formed along the circumferential direction of chamber 2. Therefore, the intake air is mixed well with fuel injected from an injection nozzle, and hence, a combustion efficiency can be improved.
Note that a ratio of a swirl speed of intake air in the cylinder chamber to a rotational frequency of an engine is called a "swirl ratio". The swirl ratio is preferably variable for the various reasons, as will be described later.
For example, when intake ports having different swirl ratios are alternately set on an identical cylinder chamber to compare engine performances, the experimental results shown in FIG. 3 are obtained. In FIG. 3, a curve indicated by a corresponds to the intake port having a high swirl ratio, a curve indicated by b corresponds to the intake port having a middle swirl ratio, and a curve indicated by c corresponds to the intake port having a low swirl ratio. As shown in FIG. 3, in the case of high swirl ratio a, when the engine speed is at low speed a1, the best engine performance can be obtained. In the case of middle swirl ratio b, when the engine speed is at middle speed b1, the best engine performance can be obtained. In the case of low swirl ratio c, when the engine speed is at high speed c1, the best engine performance can be obtained. Therefore, when the swirl ratio is constant, the performance is inevitably degraded in some engine speed range.
The reason why the engine performance varies in accordance with a change in engine speed while the swirl ratio is constant is as follows. More specifically, as shown in FIG. 4, when angle .theta. of spray flowing together with a swirl during fuel injection period .theta.inj [degree: crank angle] in unit time coincides with angle .theta.0 between two adjacent sprays, the best engine performance can be obtained. If .theta. does not reach .theta.1, this means air between adjacent sprays is not satisfactorily utilized. Therefore, if .theta. exceeds .theta.0, the immediately preceding spray overlaps the next spray, and fuel is "baked" due to air shortage in this overlapping portion. EQU .theta.0=360.degree./n (1)
[n: number of injection ports of nozzle] EQU .theta.=(Ns/Ne)(B/D).theta.inj (2) PA1 [K=constant]
[Ns: cylinder chamber air swirl speed [rpm], Ne: engine speed [rpm], (Ns/Nc): swirl ratio, B: cylinder bore diameter [mm], D: combustion chamber diameter [mm]]
For .theta.0=.theta., EQU 360.degree./n=(Ns/Ne)(B/D).theta.inj
Therefore, EQU .theta.inj.times.(Ns/Ne)=K (3)
In the relationship between injection period .theta.inj and engine speed Ne, as engine speed Ne increases, .theta.inj is prolonged, as shown in FIG. 5. Furthermore, as shown in FIG. 6, as engine speed Ne increases, the rotating speed of an injection pump is increased, and the injection pressure is increased, thereby shortening a time (injection time) required for injecting a predetermined quantity of fuel. However, as engine speed Ne increases, a crank angle in unit time (e.g., 1 ms) is also increased. As a result, as shown in FIG. 5, as the engine speed increases, the injection period is prolonged.
According to equation (3), the product of injection period .theta.inj and swirl ratio (Ns/Nc) is preferably constant K. For this purpose, as Ne increases and .theta.inj is prolonged, the swirl ratio is preferably small. Contrary to this, when the swirl ratio is constant, even if movement of a spray is matched to that of a swirl at given engine speed Ne1, .theta.inj is shortened when the engine speed is decreased below Ne1. Therefore, the swirl ratio becomes too small from equation (3), and the mixing state of fuel and air is impaired. When the engine speed is increased over Ne1, since .theta.inj is prolonged, the swirl ratio becomes too large from equation (3), and the immediately preceding spray overlaps the next spray, resulting in poor performance. In this manner, the swirl ratio is not always in proportion to the engine performance.
Engines of vehicles are subjected to the exhaust gas regulation. The amount of NO.sub.x (nitrogen oxides) produced as the major component of the exhaust gas is related to the swirl ratio, and the test results shown in FIG. 7 are obtained. More specifically, the amount of NO.sub.x produced is substantially proportional to the swirl ratio, and the higher the swirl ratio becomes, the amount of NO.sub.x produced is increased.
The engine is influenced by a load with respect to the engine speed. Therefore, the relationship between the load and the swirl ratio must be examined. In the case of low engine speed, the low swirl ratio is best suited, as described above (FIG. 3). However, as for the load, the low swirl ratio is preferable in the hatched portion shown in FIG. 8.
For example, if the load is light in the case of low speed, the high swirl ratio is not necessary and the low swirl ratio is best suited. At the middle speed, the low swirl ratio is preferable between the light load to middle load. At high speed, the low swirl ratio is best suited irrespective of a load state. More specifically, in the case of low speed and light load, this corresponds to a state of an excessive amount of intake air. Therefore, fuel can be combusted irrespective of the swirl ratio. In this state, the low swirl ratio causing a small amount of NO.sub.x produced is preferable rather than the high swirl ratio causing an increase in amount of NO.sub.x produced. As the swirl ratio is decreased, heat loss corresponding to a combustion gas absorbed by a cylinder wall is decreased. In particular, in the case of light load, the heat loss influences fuel consumption. Therefore, in this respect, the low swirl ratio is advantageous.
Conventionally, HSP structures having a low swirl ratio shown in FIG. 9A and having a high swirl ratio shown in FIG. 9B are separatly prepared. In FIGS. 9A and 9B, reference numeral 2 denotes a cylinder chamber; 6a, an intake valve seat; 5a, a low swirl intake port; and 5b, a high swirl intake port. The sectional area of high swirl intake port 5b is less than that of low swirl intake port 5a. Each intake valve seat 6a is divided into eight sections along the circumferential direction, which are denoted by numbers 1 to 8. From each number position, intake air, which propagates in a direction indicated by an arrow and has a strength corresponding to the length of the arrow, is taken in cylinder chamber 2. Intake air vectors Nos. 1 to 4 tend to swirl clockwise (+) around center 0 of cylinder chamber 2, and are called forward swirl components. Intake air vectors Nos. 5 to 8 are inevitably swirl counterclockwise (-) opposite to the forward swirl direction due to the positional relationship between intake ports 5a and 5b and cylinder chamber 2, and are called reverse swirl components. In FIG. 9A, a difference between ##EQU1## obtained by adding swirl moments Mi=Li.times.Vi (as the product of length Li of the perpendicular drawn from center 0 to each intake air vector and magnitude Vi of the intake air vector) of the forward swirl components, and that of the reverse swirl components becomes relatively small, resulting in a low swirl ratio. In contrast to this, in FIG. 9B, the total of forward swirl moments becomes large since intake port 5b is tightened, and a difference between the total of the forward swirl moments and that of the reverse swirl moments becomes large, resulting in a high swirl ratio.
A conventional structure capable of variable swirl state is described in Japanese Patent Publication No. 51-7243. This structure is as shown in FIGS. 10A and 10B. In FIGS. 10A and 10B, reference numeral 12 denotes a cylinder chamber; 15, an intake port; and 16a, an intake valve seat. Intake port 15 is constructed on the basis of the low swirl port, and divided into ports 15a and 15b by partition wall 17. Port 15b can be opened/closed by opening/closing valve 18.
When opening/closing valve 18 is opened as shown in FIG. 10A, intake air is introduced into both ports 15a and 15b, and a flow rate passing intake valve seat 16a is low. Thus, cylinder chamber 12 is set in the low swirl state. When opening/closing valve 18 is closed, as shown in FIG. 10B, intake air is guided only to one port 15a. Therefore, the sectional area of an intake air path is decreased to half and becomes smaller than the inner diameter area of intake valve seat 16a, so that the flow rate of intake air is increased. Thus, cylinder chamber 12 is set in the high swirl state. Swirl components in both the states have directions and strengths corresponding to arrows in FIGS. 10A and 10B.
In the conventional structure of this type, the swirl ratio can be varied as needed, but the following problems are posed. More specifically, in the low swirl state, as shown in FIG. 11A, a single swirl like rigid-body swirl simply occurs in cylinder chamber 12. As shown in FIG. 11B, sprays F radially injected from the center of cylinder chamber 12 into the rigid-body swirl are simply influenced by side wind of rigid-body swirl. Therefore, sprays F cannot be sufficiently mixed with air. In the high swirl state, as shown in FIG. 11C, when intake air is guided into one port 15a and passes the edge of partition wall 17, a flow path area is immediately increased. Therefore, a plurality of swirls are formed due to separation, or a loss such as reverse flow occurs. Since the flow path sectional area is decreased to half and the sectional area of port 15a is small, a large flow path resistance occurs, and intake air can flow into cylinder chamber 12 only from a part of intake valve seat 16a. Therefore, a flow rate coefficient is low, and the amount of intake air becomes short.
As the basic idea about swirl, when the high swirl state is desired, intake air is preferably introduced from the horizontal direction (circumferential direction) with respect to the cylinder chamber. In this case, the amount of intake air is small. When the low swirl state is desired, intake air is preferably introduced from the vertical direction (axial direction) with respect to the cylinder chamber. In this case, the amount of intake air is large.
However, in the conventional structure shown in FIGS. 10A and 10B, intake port 15 is simply divided into two sections, and an intake direction cannot be changed upon switching of the swirl states. Thus, in either state, the amount of intake air is undesirably decreased.
Various other conventional structures have been proposed. However, the conventional structure cannot change the swirl state while maintaining a sufficient amount of intake air, and are complicated and high in cost.